Why better environmental outcomes require us to rethink the relationship between data, decisions, and MRV.
Carbon footprints are made up.
There, I said it.
Before anyone gets too upset, I’m not suggesting that greenhouse gas emissions aren’t real, or that measuring them isn’t important. Quite the opposite.
But the carbon footprint we’ve come to treat as an objective measure of environmental performance is actually a constructed number. A useful one, perhaps, but one built on an extraordinary collection of choices, assumptions, and conventions.
Consider what goes into calculating a carbon footprint.
We choose where to draw the boundaries of a system. We decide whether to allocate emissions based on mass, economic value, or some other relationship. We select emission factors, models, and methods, each with their own assumptions and limitations. We define baselines and counterfactuals, sometimes estimating what would have happened in a world that doesn’t exist.
And then there’s nature itself.
Soil carbon, for example, is dynamic. Weather, geography, management history, and biological processes all influence what happens underground. Even with increasingly sophisticated models and measurements, isolating the effect of a particular management decision is difficult.
Perhaps most revealingly, we can change a carbon footprint without changing anything in the physical world.
Swap a generic emission factor for a more representative one. Adopt a different model. Replace industry averages with primary data. The reported number might improve, or it might get worse, even though the underlying activity hasn’t changed.
None of this means carbon accounting is fraudulent. It means it’s accounting.
And accounting, by definition, requires conventions.
So why bother?
Because without a shared accounting system, we’d struggle to establish common expectations, compare performance, direct investment, or hold anyone accountable.
Financial accounting is hardly a perfect representation of economic reality. Depreciation, asset valuation, and revenue recognition all involve conventions and judgments. Yet we depend on those conventions to coordinate decisions across an extraordinarily complex economy.
Carbon accounting serves a similar purpose. Common standards help us make sense of a complicated world.
But here’s where I think we’ve gotten something backward.
We’ve become so focused on measuring environmental performance that we’ve sometimes lost sight of the decisions the measurements are supposed to inform.
The goal was never to produce a better carbon footprint.
The goal was to produce a better environmental outcome.
Those are not necessarily the same thing.
The difference between better accounting and better decisions
Consider a company trying to reduce its agricultural supply chain emissions.
One approach is to build a more sophisticated carbon inventory. Collect better farm-level data. Refine the emission factors. Improve the models. Identify emissions hotspots and report progress against a target.
There’s real value in that work.
But imagine instead that we started with a different question:
What would actually need to change for this agricultural system to become more sustainable?
Suddenly, the conversation expands.
A farmer might need to change crop rotations, adopt different management practices, or invest in more efficient equipment. But those decisions depend on profitability, agronomic risk, market access, financing, infrastructure, and the incentives offered by buyers.
A food company might want lower-emission ingredients, but its procurement requirements could make it difficult for farmers to change what they grow.
A lender might support investments in more resilient production systems, but struggle to evaluate whether those investments reduce financial risk.
And in a water-constrained region, a practice that reduces emissions might do little to address groundwater depletion, or could even create unintended trade-offs.
These aren’t primarily carbon accounting problems.
They’re systems decision problems.
They require understanding how different actors, incentives, constraints, and outcomes interact.
And they require data that helps people evaluate their options, not simply document what they’ve already done.
Better data, but for what?
There’s a familiar refrain in sustainability: we need better data.
I agree. But I’d add a qualification.
We need better data for better decisions.
Not every decision requires the most precise measurement. Not every uncertainty needs to be resolved before action is possible. And not every additional data point improves our understanding of what to do next.
A farmer deciding whether to grow a different crop might care far more about expected profitability, yield variability, irrigation requirements, and market access than a highly refined estimate of soil carbon change.
A company considering an investment in regenerative agriculture might need to know which barriers prevent adoption before it needs another decimal place on an emissions factor.
A financial institution evaluating a water conservation investment might need to understand whether reduced irrigation translates into actual reductions in aquifer depletion, not merely whether pumping becomes more efficient.
The right data depends on the decision.
This also changes how we think about uncertainty. Rather than treating uncertainty as something to eliminate at any cost, we can ask: Which uncertainties could actually change the decision?
That question helps us distinguish between information that’s interesting, information that’s reportable, and information that’s actionable.
It also changes how we approach sustainability at a landscape or systems level.
An individual farm may optimize its own water use, but that doesn’t necessarily translate into improved aquifer conditions. A buyer may reduce its reported Scope 3 emissions without fundamentally changing production practices. A project may generate verified environmental credits without creating a durable business case for the practices behind them.
Understanding these differences requires more than a better footprint.
It requires understanding the system itself, including the relationships between decisions made by farmers, buyers, lenders, policymakers, and other participants.
What if MRV were the byproduct, not the starting point?
This brings me to something I’ve been thinking about increasingly in my work.
We’ve invested enormous effort into measurement, reporting, and verification (MRV). We’ve built methodologies, protocols, platforms, registries, and increasingly sophisticated tools for calculating environmental outcomes.
Much of this infrastructure is necessary.
But too often, we begin with the MRV framework and then try to build a functioning program around it.
We decide what needs to be measured, design the reporting requirements, and ask participants to supply the information. Then we hope that the incentives attached to those measurements will produce meaningful change.
What if we reversed the sequence?
What if we started by understanding the decisions that need to happen, the conditions that make those decisions possible, and the information people need to act?
Imagine a system where farmers, buyers, financial institutions, and other participants can evaluate different management options using a shared understanding of the relevant data.
A farmer can assess the economic and agronomic implications of changing practices. A buyer can understand what procurement commitments might enable those changes. A lender can evaluate the financial risks and opportunities. Other stakeholders can assess the likely consequences for water, carbon, and other environmental outcomes.
As decisions are made and activities take place, the system captures relevant evidence through normal operations, supplemented by targeted measurements where necessary.
That evidence can then support environmental accounting, reporting, and verification using transparent, interoperable methods.
In this paradigm, the world once described by my friend Dorn Cox comes to fruition:
MRV becomes an output of a functioning decision-making system, rather than the primary reason the system exists.
Importantly, this doesn’t mean verification is automatic or that measurement requirements disappear. Independent verification, scientific rigor, and appropriate data collection remain essential.
But it does mean we design data systems around creating value for the people making decisions, while ensuring that credible environmental evidence can be produced from those systems.
The same underlying information can support different purposes without requiring everyone to rebuild the data collection process from scratch.
This is where the potential for harmonization becomes especially interesting.
Harmonize the evidence, not necessarily every answer
The instinct in environmental accounting has often been to drive consistency through standardized models, methods, and reporting requirements.
And consistency matters.
But different decisions legitimately require different approaches.
A corporate greenhouse gas inventory, an agricultural investment decision, a water conservation program, and a carbon credit project aren’t trying to answer exactly the same question.
They shouldn’t necessarily use the same model, the same system boundary, or the same performance metric.
What they should have is a shared foundation of trustworthy information.
That means clearer definitions, transparent assumptions, traceable data sources, interoperable formats, appropriate permissions, and a way to understand how different methods translate the same underlying evidence into different outputs.
Rather than forcing every participant into one accounting framework, we can make it easier for different frameworks to use common evidence while being transparent about why their answers differ.
This is particularly important in agriculture, where the same parcel of land may be involved in supply chain reporting, conservation programs, financing decisions, regulatory requirements, and environmental markets.
The farmer shouldn’t have to become an expert in five different accounting systems just to participate.
Nor should a buyer or investor need to create an entirely new data collection infrastructure every time it wants to evaluate a different outcome.
A shared evidence foundation could reduce that friction while making environmental accounting more transparent and useful.
And perhaps more importantly, it could help us understand whether the decisions we’re making are actually working.
This fragmentation is part of what led me to build SADDLE, an effort to map the relationships between sustainability tools, the decisions they support, the data they require, and the methods behind them.
The more I’ve explored this landscape, the clearer it’s become that we have an abundance of tools for measuring and reporting environmental performance, but a much less coherent system for helping people decide what to do.
The challenge isn’t simply connecting more data to more tools. It’s understanding which information is actually useful for a particular decision—and making sure we can learn from what happens next.
From measuring performance to enabling change
None of this is an argument for abandoning carbon accounting.
We need it. We need credible measurement, shared rules, independent verification, and accountability.
But we also need to recognize its limitations.
A carbon footprint can tell us something about the environmental performance attributed to an activity under a particular set of rules. It cannot, on its own, tell us what decision to make, what barriers need to be overcome, or whether an intervention will create lasting change.
Those questions require a broader understanding of the systems we’re trying to influence.
And this is where I see an opportunity to rethink the role of environmental data.
Instead of asking how we can collect more information to satisfy reporting requirements, we might ask:
What decisions are we trying to enable? Who needs to make them? What information would change their choices? And how can we capture credible evidence of what happens as a result?
That doesn’t make accounting less important. It gives accounting a more useful place in the process.
We should absolutely continue improving our ability to measure environmental outcomes. But the more important measure of progress is whether those improvements help people take actions that otherwise wouldn’t happen.
After all, the atmosphere doesn’t care which emission factor we used. An aquifer doesn’t care how sophisticated our reporting platform is.
What matters is what actually changes.
The future of environmental accounting shouldn’t just be better measurement of the world as it is. It should be part of a system that helps us make the world better, and tells us whether we’re succeeding.